• 한국해양공학회지
• /
• 제35권1호
• /
• pp.91-97
• /
• 2021

Marine accidents caused by ships have brought about economic and social losses as well as human casualties. Most of these accidents are caused by small and medium-sized ships and are due to their poor conditions and insufficient equipment compared with larger vessels. Measures are quickly needed to improve the conditions. This paper discusses a video-integrated collision prediction and fall detection system to support the safe navigation of small- and medium-sized ships. The system predicts the collision of ships and detects falls by crew members using the CCTV, displays the analyzed integrated information using automatic identification system (AIS) messages, and provides alerts for the risks identified. The design consists of an object recognition algorithm, interface module, integrated display module, collision prediction and fall detection module, and an alarm management module. For the basic research, we implemented a deep learning algorithm to recognize the ship and crew from images, and an interface module to manage messages from AIS. To verify the implemented algorithm, we conducted tests using 120 images. Object recognition performance is calculated as mAP by comparing the pre-defined object with the object recognized through the algorithms. As results, the object recognition performance of the ship and the crew were approximately 50.44 mAP and 46.76 mAP each. The interface module showed that messages from the installed AIS were accurately converted according to the international standard. Therefore, we implemented an object recognition algorithm and interface module in the designed collision prediction and fall detection system and validated their usability with testing.

• 한국컴퓨터정보학회논문지
• /
• 제25권1호
• /
• pp.45-53
• /
• 2020

기존의 CNN 알고리즘은 위성영상과 같은 대형 이미지에서 소형 객체를 식별하는 것이 불가능하다는 문제점을 가지고 있었다. 본 연구에서는 이러한 문제를 해결하기 위해 관심영역 설정 및 이미지 분할 기법을 적용한 CNN 알고리즘 개선방안을 제시하였다. 실험은 비행장 및 항공기 데이터셋으로 전환학습한 YOLOv3 / Faster R-CNN 알고리즘과 테스트용 대형 이미지를 이용하여 진행하였으며, 우선 대형 이미지에서 관심영역을 식별하고 이를 순차적으로 분할해 나가며 CNN 알고리즘의 객체식별 결과를 비교하였다. 분할 이미지의 크기는 실험을 통해 최소 분할로 최대의 식별률을 얻을 수 있는 최적의 이미지 조각 크기를 도출하여 적용하였다. 실험 결과, 본 연구에서 제시한 방안을 통해 CNN 알고리즘으로 대형 이미지에서의 소형 객체를 식별하는 것이 충분히 가능함을 검증하였다.

• 대한물리의학회지
• /
• 제9권3호
• /
• pp.243-248
• /
• 2014

PURPOSE: This study was aim to the change of muscle activities of lower extremity and waist during lifting a small object on the floor according to different foot position of women in their twenties wearing a skirt. METHODS: 9 women in their twenties wearing a skirt were selected and were measured the muscle activities of medial gastrocnemius (MG), tibialis anterior (TA), vastus lateralis (VL) and iliocostalis (IC) when they lifted a small object on the floor. The two different foot position employed were "both feet posed straight side by side" (condition 1) and "both feet posed diagonally to 45 degree" (condition 2) used. The order of feet position was selected randomly and the subject took a rest for 30 min between tests to prevent muscle fatigue. We calculated the mean and standard deviation of muscle activities and used Mann-Whitney U test to compare the difference between the two foot positions with SPSS(IBM Korea) RESULTS: The muscle activity of condition 2 was greater than that of condition 1 in right side of TA, VL, and IC and left side of TA, VL, MG and IC. The right side of TA, VL and left side VL were significant difference between condition 1 and condition 2(p<.05). CONCLUSION: We suggest "both feet posed straight side by side" position is better if a woman wearing a skirt lift the small object and it will help prevent the low back and lower limb problems in the future.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1995년도 춘계학술대회 논문집
• /
• pp.408-413
• /
• 1995

Nowadays, manufacturing system is trending toward integrated circumstance by helping of CAD/CAM/CAT. To use this system effectively, it is necessary to get exact 3 dimensional surface data of an object. We have been using contact method to measure 3 dimensional object profile. But his method has demerit of leaving scrach or small distortion on the object. To improve this, a non-contact measuring method using optical system is adopted. In this paper, We propose optical ring method. Withthis system, We could measure displacement of the object in the range of 45mm having 150mm having 150 .mu. m resolution with no scratch or distortion.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2002년도 학술대회 논문집 전문대학교육위원
• /
• pp.83-87
• /
• 2002

Accurate acquisition of surface geometries such as machined surfaces, biological surfaces, and deformed parts and processing 3D object have been very important technique in scientific study and engineering, expecially for system design, manufacturing and inspection. Defective human teeth are usually ground to be coped with special alloy coping which is hand-made by dental technician. This make dental technician to be difficult and take a long time Dental CAD/CAM Systems consist of two parts, data acquisition and milling. In this paper, a method is studied to mill object which is acquired 3D geometric data of the small object such as a die in stone model. This paper present a control program and a mechanical system for milling 3D object.

• 한국정밀공학회지
• /
• 제16권9호
• /
• pp.166-171
• /
• 1999

An overhead crane system consists of trolley, girder, rope, object, trolley motor, girder motor, and hoist motor. An analytic model which derived from the modelling of the overhead crane system is nonlinear model which includes the swing and the twist angle of the object. this model consists of the equation of motion for motors and object. If the swing angle and the acceleration of Z for the object are small, this model becomes a simple nonlinear model which doesn't include the swing and the twist angle of the object. From the results of computer simulation, the characteristics of an actual overhead crane system could be predicted by the simple nonlinear model.

• 로봇학회논문지
• /
• 제12권1호
• /
• pp.33-41
• /
• 2017

We present a region-based approach for accurate pose estimation of small mechanical components. Our algorithm consists of two key phases: Multi-view object co-segmentation and pose estimation. In the first phase, we explain an automatic method to extract binary masks of a target object captured from multiple viewpoints. For initialization, we assume the target object is bounded by the convex volume of interest defined by a few user inputs. The co-segmented target object shares the same geometric representation in space, and has distinctive color models from those of the backgrounds. In the second phase, we retrieve a 3D model instance with correct upright orientation, and estimate a relative pose of the object observed from images. Our energy function, combining region and boundary terms for the proposed measures, maximizes the overlapping regions and boundaries between the multi-view co-segmentations and projected masks of the reference model. Based on high-quality co-segmentations consistent across all different viewpoints, our final results are accurate model indices and pose parameters of the extracted object. We demonstrate the effectiveness of the proposed method using various examples.

• 한국정밀공학회지
• /
• 제21권3호
• /
• pp.51-58
• /
• 2004

This paper describes the development of a small 6-axis force/moment sensor for robot's finger, which measures farces Fx, Fy, Fz, and moments Mx, My, Mz simultaneously. In order to safely grasp an unknown object using the robot's gripper, and accurately perceive the position of it in the gripper, it should measure the force in the gripping direction, the force in the gravity direction and the moments each direction, and perform the force control using the measured forces and moments. Also, it should detect the moments Mx (x-direction moment), My and Mz to accurately perceive the position of the object in the grippers. Thus, the robot's gripper should be composed of 6-axis force/moment sensor that can measure forces Fx, Fy, Fz, and moments Mx, My, Mz simultaneously. In this paper, the small 6-axis force/moment sensor for measuring forces Fx, Fy, Fz, and moments Mx, My, Mz simultaneously was newly modeled using several parallel-plate beams, designed, and fabricated. The characteristic test for the developed sensor was performed, and the result shows that intereference errors of the developed sensor are less than 4.23％. Thus, the developed small 6-axis force/moment sensor may be used a robot's gripper.

• 로봇학회논문지
• /
• 제19권1호
• /
• pp.58-64
• /
• 2024

The ability of the robot gripper to handle a wide range of objects significantly impacts its operational effectiveness. Among the robot grippers commonly used, the economically feasible choice is the relatively simple structure of a parallel gripper. To perform more densely packed tasks with a parallel gripper, it should be capable of handling small objects. Therefore, this study designs a parallel gripper mechanism equipped with assisting grippers to ensure smooth grasping of small objects. The parallel gripper is designed using a rack and pinion gear system, with two additional grippers on both side, and these assisting grippers are designed to be detachable. The two assisting grippers have different type of tip to grasp thin fabric shapes and thin stick shapes. The gripper prototype is used to verify the grasping capabilities for shapes achievable with a conventional parallel gripper and those intended for grasping with the assisting grippers through grasping experiments. Consequently, by equipping a conventional parallel gripper with assisting grippers as in this study, it becomes capable of handling a broader range of objects, in addition to its existing functionality.

• 한국측량학회지
• /
• 제11권2호
• /
• pp.7-15
• /
• 1993

대상 지역의 자료기반을 구축하기 위해서는 필요한 목적에 따라 신속하고 정확한 지형정보를 얻어야 한다. 일반적으로 지형정보의 취득에 이용되는 방법에는 지도, 위성영상 및 항공사진의 입체모형을 이용하는 방법이 있으나 대상지역의 크기, 영상의 공간해상력, 소요정확도 및 경제성 등을 고려하여 최적방법을 선택하여야 한다. 따라서, 본 연구에서는 소규모 대상지역인 농촌 정주권사업 계획 지구를 선정하여 Photoballoon 시스템을 활용한 지형정보의 취득 및 기본도제작 방법과 지형공간정보체계의 공간 중첩기법을 활용함으로써 소규모 대상지역의 효율적인 지형자료기반 구축방법을 제시하는데 목적이 있다. 본 연구의 수행결과, 대상지역의 정밀한 입체사진을 얻기 위한 Photoballoon 시스템의 최적 촬영조건을 결정하고 이에 따른 정확도 분석을 통해 소규모지역에 대한 입체사진 및 지형도를 신속하고 경제적으로 얻을 수 있었으며, 이를 이용하여 대상지역에 필요한 각종 계획을 수행함으로써 소규모지역의 효율적인 계획 및 설계에 활용가능한 자료기반을 구축할 수 있었다.